Manus’s more than $500 million raise does not prove its agents are changing work

Neither the funding nor the reported valuation shows how many people use Manus for paid work, which tasks they use it for, or whether their hours, pay or working conditions have changed. China’s National Development and Reform Commission prohibited Meta’s acquisition of Manus and required the parties to withdraw, The Associated Press reported. Manus later said it had resumed independent operations on Sept. 1. That corporate history helps explain the context for the raise; it does not establish what the software has done for workers.

What can Manus agents do—and what remains unproven?

Manus describes its agents as tools for computer-based tasks. Its product materials say Manus 2.0 is available, with automations linked to events in connected services, a shared workspace and computer-use features. The company describes Cue, a separate agent app, as being in early access. These are product descriptions, not independent evidence of how well the tools perform in workplaces. Manus’s materials also cover creative work and transactions.

Meta said in December 2025 that Manus served millions of users and businesses worldwide. It also reported more than 147 trillion tokens and more than 80 million virtual computers, figures reproduced by PYMNTS. These company-reported activity measures are not independently audited counts of workplace use. They do not show how many people rely on Manus for work or whether the product has changed employment, pay or productivity.

A possible case for agents is that automating computer-based tasks could save people time. But that remains an outcome to test, not a conclusion that follows from a funding announcement. SiliconANGLE reported Manus’s claim that Manus 2.0 uses 23.2% fewer tokens and processes tasks more than 28% faster in some cases; those performance figures have not been independently verified. Even a faster task or lower token use does not show whether workers save time overall, get better results or have better jobs.

The usability study offers a reason to test those outcomes rather than assume them. Its 31 participants assessed Manus and Operator on selected tasks; the researchers reported friction involving mismatched expectations, trust, communication and collaboration. The findings do not represent every user or workplace. They do make task-specific evaluation more useful than relying on user totals or technical performance claims.

What do users risk when a company’s plans change?

Workplace effects are not the only human stakes. Manus said its separation from Meta required some users to back up and restore data, and that some experienced temporary interruptions in access. The company also said certain users’ data generated on or after Dec. 29, 2025, would be deleted during a specified August 2026 window to meet regulatory requirements in some jurisdictions. Manus’s user notice and its update on resuming independent operations describe changes that affected users directly.

That transition does not prove workplace harm. It does show why people need clear notice, practical ways to retrieve their data and clarity about who is responsible when an agent or a service change disrupts a task. For software connected to work accounts and workflows, reliability and recourse matter alongside speed.

Employers and investors should demand evidence

The strongest case for Manus’s backers is that capital can support tools that may automate routine digital tasks. But neither a large raise nor company-reported usage establishes that workers benefit. Evidence should track actual deployments: which tasks agents handle, how often their output needs correction, how much human review is required, and whether workers’ time, workload or job conditions improve or worsen.

Employers should require independent, task-specific testing before assigning agents consequential work, with human review and a clearly identified person responsible for correcting mistakes. They should report deployment and outcome measures in terms workers can understand, not just system activity. Investors can make transparent reporting a condition of their support, and Manus can disclose how its tools perform and what recourse users have. The announced raise is evidence of investor commitment, not evidence of workplace outcomes.